A study on simultaneous extraction of SNS communities and terms

2017 Fiscal Year Final Research Report

• Project/Area Number: 15K00309
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Intelligent informatics
• Research Institution: The University of Tokushima
• Principal Investigator: Yoshida Minoru 徳島大学, 大学院社会産業理工学研究部(理工学域), 講師 (40361688)
• Project Period (FY): 2015-10-21 – 2018-03-31
• Keywords: SNS / 分散表現 / コミュニティ抽出

Outline of Final Research Achievements

We studied a method for extracting communities (i.e., the users that have similar characteristics) and the terms used in the communities simultaneously, based on the assumption that similar communities use similar terms. We focus on the profiles written by the users on Twitter, and we developed a system that extract the frequently-used terms by the users with the profiles that include the given query word. We also developed a method for word distributed representations that reflect the co-occurrence of the words in profiles and tweets, and cluster users and tweets based on the learned vector representation.

Free Research Field

自然言語処理